Nuclear disasters are so rare that a single data point of the magnitude of damage provides so much information on our risk assessment of nuclear power that it seems reasonable to idle plants while we wait and see how this one plays out. This is only the third major issue to ever happen.

Waiting this one out doesn't tell us more about the frequency of occurrence (it already occurred), but it does tell us about the magnitude of damage, which is one of the biggest unknowns and arguably why the US caps nuclear plant liability at ~12 billion via http://en.wikipedia.org/wiki/Price%E2%80%93Anderson_Nuclear_... .

The overt reason was that it was such an unknown that no one could insure it; a cynic might say that the cap was known to be incredibly low from the outset, and it was really just a hidden subsidy towards the creation of fuel for more and more nuclear weapons (this was the late 50s).